Environmental Engineering Reference
In-Depth Information
and [TRA 94]. They found that the size of a representative specimen for the
measurement of hydraulic conductivity of compacted soil liners depends on the
method and quality of construction. If the soil is poorly compacted, then the
representative specimen size should be very large. However, when the soil is well
compacted, the representative size is small and close or equal to the dimension of
standard laboratory test specimens.
The above results and related observations have recently been confirmed (Figure
16.11) by field data presented by [DAN 97, DAN 98]. The database considered was
assembled from information from more than 120 sites. In order to be included, the
compacted clay liners (CCLs) had to be [DAN 89, TRA 94]:
− constructed in accordance with industry practice for a full size liner;
− subjected to a quality assurance program that was in accordance with industrial
practice;
− constructed with the objective of demonstrating that large-scale hydraulic
conductivity (
K
field
) did not exceed 10
-9
m/s;
− reasonably well documented test results; and
− availability of results from large-scale field hydraulic conductivity tests, such
as the sealed double ring infiltrometer.
The data presented in Figure 16.11 are in full agreement with the range of
variation of
in situ
hydraulic conductivity (
k
field
)
versus
thickness of compacted clay
liners reported by Benson and Daniel [BEN 94a, BEN 94b]. Figure 16.11 clearly
indicates that hydraulic conductivity decreases with increasing thickness of the clay
liner. It is worth noting that most of the CCLs investigated lay in the region where
the construction of the liner was deemed to be good to excellent.
The index properties of the CCL considered together with the limit values
suggested by Jones
et al.
[JON 93] and Daniel [DAN 89] in order to succeed in
obtaining a low hydraulic conductivity at the field scale are given in Figure 16.12,
together with some data from Italian and Australian sites for comparison purposes.
Daniel's database [DAN 97] shows that there the index properties have little
influence on
k
field
. Other factors appear to be far more important for complying with
k
field
<10
-9
m/s requirement.
Considering that 75% of the CCLs achieved the objective of
k
field
<10
-9
m/s and
that 25% of the total did not, it is interesting to show and comment on the results
reported in Figures 16.10 and 16.13.
